基于pca和网格特征的车牌识别算法分析license plate recognition algorithm based on pca and the grid feature analysis.docx

摘要
车牌识别系统是智能交通系统中的关键部分,也是国内外学者研究的热点。它结合了图像处理及模式识别等多项技术,可以广泛应用于高速公路自动收费、道路监控、小区及停车场车辆管理系统等领域。论文主要对车牌字符分割及字符识别进行了深入地研究,提出了相关改进算法,并利用 VC++ 平台开发出了论文中的车牌识别系统。
论文分析了已有的各种车牌识别算法,提出了改进的车牌识别算法。对于车牌图像的倾斜校正,论文采用了自适应阈值的 Canny 边缘检测算法,能够检测到完整的边缘信息,从而可以准确地确定出车牌的倾斜角度,实现对车牌的校正。对于车牌区域的二值化,论文给出了自适应步长的二值化算法,该算法抗噪能力
好,自适应性强。在此基础上,论文通过基于垂直投影的方法进行字符分割,有效地提高了字符分割的成功率。对于字符识别,论文针对车牌字符的特点,对汉字采用了基于 PCA(主分量分析)特征的多模板匹配算法;对字母数字则采用了基于网格特征的单模板匹配算法。
最后,论文对车牌识别系统进行了相关测试。测试结果表明,该系统能有效地实现车牌定位、字符分割及识别,系统的整体识别率达到 %。
关键字:字符分割二值化字符识别 PCA 特征网格特征
Abstract
License Plate Recognition System is one of the most critical part in the Inte lligent Transportation Systems, is also the research fouse at home and abroad. It contains many technologies, including image processing, pattern recognition and so on, which can be widely used in the field of automated highway toll collection, traffic monitoring, community and parking management system. The license plate character segmentation and characters recognition are studied deeply in this paper, and puts forward some improved algorithms, and by using VC++ platform to develop license plate recognition system in the paper.
The existing license plate recognition algorithms are analysed and improved
algorithms are given by the paper. For license plate image tilt correction, an adaptive threshold Canny edge detection algorithm is used. The method is able to detect plete edge information, which can accurately determine the angle of the inclined plate, to realize the calibration of the license plate. For the license plate area binarization, an adaptive step binarization algorithm is proposed. It has an ability of resisting noise and a strong adaptability. O n this basis, a method based on vertical projection is used to segment characters, effectively improves the ess rate of the character segmentation. For character recognition, aiming at the characteristics of license plate characters, for chinese characters, a multi-template matching algorithm based on PCA(